資料命名網路架構下之車載通訊緊急應用

Abstract

命名中心網路(Named Data Networking, NDN)主要是以資料的名稱來命名而非現今網路的IP位址來進行資料的傳輸，他與現今封包格式最大的不同在於他所著重的地方是資料的名稱而非資料的位置。在車載網路(Vehicle Network)當中，訊息的流動大多都夾雜著像是車禍訊息及道路車流量..等，在這大量訊息流動的網路之中，使用者所考慮的內容幾乎是該訊息的名稱而非位置，這種特性也與命名中心網路的性質非常類似。 在此篇論文當中，我們結合了命名中心網路與車載網路的網路架構，提出了一個能夠讓車與車間自主溝通的系統。在此系統中，訊息的名稱主要是以發送者(車輛)的位置、預定興趣表(Pending Interest Table)中的興趣(Interest packet)為組成的主旨，讓周遭的接收者能夠迅速的判斷該訊息是否符合自身的需求並納入內容儲存庫(Content Store)之中。在此次的實驗之中，我們會去改變傳統命名中心網路的預定興趣表、內容儲存庫以及與上下層連接的介面，使其能夠符合在緊急情況下須要主動推播資訊的特性，發揮命名中心網路的功能迅速的讓周遭的車輛收到緊急訊息。 依照不同類型的應用程式，生成的封包名稱也會有些許的不同，依據不同的需求，在PIT及CS做不同的事情，不管要發送或是接收任何類型的訊息，都能採取最有效的策略去進行。 最後，透過我們的模擬及數據可以顯示出 eVNDN 擁有著 1) 對於各種緊急情形下更為優秀的訊息傳遞成功率 2) 更為廣闊的傳遞範圍 3) 比起傳統車載網路更為短暫的延遲時間，並利用這些實驗結果指出未來能夠改善以及研究發展的方向。

Unlike current network, Named Data Networking (NDN) focuses on “what” is the data instead of “where” is the data. The major difference between NDN and traditional packet network is that the routing is based is the name of the contents, not the location (i.e., IP address) of the content. In vehicular network, warning message usually included car accidents, road traffic flows, etc. When large amount of messages flow through the vehicular network, users only interest the content in the messages that are related to them. Thus, this characteristic is very similar to the NDN. In this thesis, we proposed a system called Emergency Application for Vehicle-to-Vehicle Communication using Named Data Networking (eVNDN). This enables vehicles to communicate based on the contents of the messages autonomously. In eVNDN, the name of the message is combined with the location of the senders (e.g., vehicles) and the interests of the Pending Interest Table (PIT). This allows nearby neighbor vehicles quickly determine whether these messages fit for their interest. If so, those messages will be included into the Content Store (CS). In eVNDN, we changed the PIT, CS, and both upper and bottom connecting interfaces of NDN. This allows users (e.g., vehicles) automatically send and receive messages related to emergency applications with the functions of NDN. In eVNDN, generated packet names are based on different emergency broadcast applications. According to different demands and applications, eVNDN will process the received packets differently (e.g., cache and drop) in PIT and CS. Based on the content in the packets that users (e.g., vehicles) send or receive, eVNDN will determine the optimal strategy. Finally, our experiments and simulation results to showed that eVNDN have 1) high successful delivery rate of messages for different kinds of emergency application; 2) longer range of the emergency messages can be sent and; 3) shorter delay for emergency messages in the vehicular network. With the experiment results, we showed that the eVNDN is able to improve the performance of the vehicular network for emergency applications.